Regression Models for Categorical Dependent Variables Using Stata

By J. Guzmán. 2013

4,049 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/e5a7d7e06f621cc1af2adb9a9f2db20dd5e7fa01

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

National Animosity and Cross-Border Alliances (2013, methodology): “…embedded in a network and tracked over multiple years, we used the Stata code for the three-way clustering of standard error terms in nonlinear estimations of Poisson, rare-events logit, negative binomial, ZINB (Cameron & Tivendi, 1998; Long & Freeze, 2001), and zero-inflated-Poisson models.”
Don't Just “Like” Me, Promote Me: How Attachment and Attitude Influence Brand Related Behaviors on Social Media (2018, methodology): “Therefore, we conducted tests for overdispersion and the results revealed that negative binomial regression was the more appropriate analytical procedure (Long and Freese 2006).”
Low cost private schooling in India: Is it pro poor and equitable? (2011, methodology): “Usual regression analysis techniques assume absolute independence of each observation in the dataset (Long and Freese, 2006), so the violation of this assumption of independence is taken into account through the clustering variable (the household number) using the robust cluster command, yielding…”
Defining reintroduction success using IUCN criteria for threatened species: a demographic assessment (2015, methodology): “We used generalized linear models (GLMs), hierarchical partitioning (HP, Chevan & Sutherland, 1991) and ordered logit models (OLMs, Long & Freese, 2001) to examine the relationships between ecological input parameters, extinction probabilities and status.”
A Marginal Effects Approach to Interpreting Main Effects and Moderation (2021, background): “Over the past couple of decades, researchers have become interested in how the general distributions of variables may influence outcomes (Bowen, 2012; Long & Freese, 2014; Wiersema & Bowen, 2009).”
Crime, fear and subjective well-being: How victimization and street crime affect fear and life satisfaction (2013, methodology): “Ordered logit models (STATA’s ologit), however, could not be used because the Brant test showed that some variables violated the assumption of parallel regression (Long and Freese, 2003: 165–8).”
Informational Lobbying in the European Union: The Effect of Organisational Characteristics (2012, methodology): “Applying a linear regression model to this outcome variable could result in inefficient, inconsistent and biased estimates (Long and Freese 2006 : 349).”
Competing Explanations of Undergraduate Noncompletion (2011, methodology): “These tests did not compare the sheaf coefficients themselves across types of college but instead undertook a joint Wald test (Long & Freese, 2006) on the coefficients of the variables that Undergraduate Noncompletion underlie each sheaf.”
Regional competitiveness, university spillovers, and entrepreneurial activity (2011, unknown): “Nevertheless the empirical models presented in the following subsections were tested against alternative models using the ‘countfit’ procedure suggested by Long and Freese (2006, p. 409) and found the negative binomial regression model superior to the Poisson regression, zero inflated Poisson regr”
Self-selection and variations in the laboratory measurement of other-regarding preferences across subject pools: evidence from one college student and two adult samples (2013, methodology): “We tested the IIA assumption using the two tests presented by Long and Freese (2006), the Hausman test and the Small-Hsiao test.”
Distinct spatial characteristics of industrial and public research collaborations: evidence from the fifth EU Framework Programme (2010, background): “This problem is solved by introducing a stochastic heterogeneity parameter exp(ξ ij ), leading to a negative binomial density distribution that is given by Long and Freese (2001)”
Cognitive Biases in Performance Evaluations (2016, methodology): “When the assumption is violated, it is not advisable to use an ordinary least squares model instead because the assumptions behind applying OLS to ordinal data are even stronger (Long and Freese 2014, pp.326–331).”
Determinants of the Adoption of Climate-Smart Agricultural Practices by Small-Scale Farming Households in King Cetshwayo District Municipality, South Africa (2019, methodology): “Generalized ordered logit regression can estimate models with lesser restrictions than the proportional odds or parallel lines models estimated by the ordinary ordered logit regression, whose assumptions are often violated, but more parsimonious and interpretable than those predicted by non-ordinal ”
To Merge, Sell, or Liquidate? Socioemotional Wealth, Family Control, and the Choice of Business Exit (2019, unknown): “In each analysis, the omitted alternative becomes the base outcome (Hole, 2007; Long & Freese, 2006).3 Dependent Variable In the mixed logit analysis after matching, the dependent variable contains 1 for the selected alternative (continuation, sale, dissolution, or merger) and 0 for nonselected…”
Sample Selection Bias and Heckman Models in Strategic Management Research (2016, methodology): “Much has been written about model fit in probit models (e.g., Hosmer, Hosmer, Le Cessie, and Lemeshow, 1997; Long and Freese, 2006), particularly about pseudo-R2 as a metric.”
The Dimensions of Public Service Motivation and Sector Work Preferences (2012, background): “Incidentally, this is why the model is sometimes referred to as the exploded logit model, since 1 observation with ranks is ‘exploded’ into multiple observations, with one observation for each rank” (Long & Freese, 2006, p. 342).”